RU194496U1 - ADAPTIVE DIGITAL FILTER FOR THE SUPPRESSION OF NON-FLUCTUATION INTERFERENCE - Google Patents

Abstract

The utility model relates to radio engineering and can be used in digital signal processing devices, with the possibility of distortion in the channels of external non-fluctuation interference. The technical result consists in increasing the accuracy of filtering non-fluctuation noise and in reducing the level of non-fluctuation noise at the output of an adaptive digital filter. The adaptive digital filter contains the first and second delay registers, multipliers, adders, integrators, an adaptive weighting adjustment circuit, containing a derivative calculation unit and a module calculation unit connected in series, the first and second delay registers, the input of the latter is connected to the output of the complex envelope signal converter and complex conjugate envelope.

Description

The technical field to which the utility model belongs.

The utility model relates to the field of radio engineering and can be used in digital signal processing devices passing through communication channels, in which there is the possibility of signal distortion associated with the presence of external non-fluctuation interference in the channels.

The level of technology.

As you know, interference is any effect superimposed on a useful signal and making it difficult to receive. The interference is very diverse both in its origin and in physical properties [Electronic resource: https://koralexand.ru/?page_id=105].

At the input of the receiver, along with fluctuation (noise) interference, various non-fluctuation (non-noise, structural) interferences are often present, the most typical of which are the following: harmonic signal, signal with binary phase shift pseudo-random sequence (PSP-FM), relayed signal [Electronic resource : http://www.autex.spb.su/download/dsp/dspa/dspa2003/tom1_23.pdf].

An effective way to combat non-fluctuation noise in communication channels is to use adaptive filters (AF), which are usually implemented as non-recursive digital filters with adjustable weight coefficients (VC), included at the input of the signal demodulator.

An element of the output sequence of a non-recursive filter of length N in the space of complex envelopes can be written as follows:

where X _i = [x _i , x _i-1 , ... x _{i-N + 1} ] ^T is the vector of the sequence of input samples;

- вектор ВК в i-й момент времени; т - символ транспонирования.

- VC vector at the i-th moment of time; t is the transpose symbol.

With the same structure, filters of this type differ in the ways of tuning the VC vector, aimed at minimizing a given objective function.

From the prior art [copyright certificate SU 1707740 A1, publ. 01/23/1992] a digital non-recursive filter is known, which includes a delay register, seventeen delay recorder blocks, sixteen adders and a computer connected to the corresponding binary switching.

A disadvantage of the known device is that it does not have the ability to adapt to signal parameters, that is, the filter uses the initially set signal correction parameters that are not dependent on changes in interference parameters.

The prior art [patent RU 99623 U1, publ. November 20, 2010] an adaptive signal filter is known, consisting of an input device block, an adaptation loop block, a dynamic error compensator block, and an identifier block. The blocks of this device, by means of the electric and / or electronic elements included in them, implement the principle of filtering a scalar message with a random noise dispersion identification block, which allows you to compensate for the a priori unknown noise stream of the processed signal caused by an a priori unknown flow of measurement errors and / or source state error signal, and / or external noise.

A disadvantage of the known device is that it filters only noise interference, and filtering non-fluctuation interference is impossible in it.

From the prior art [copyright certificate SU 1764145 A1, publ. 09/23/1992] an adaptive filter is known that contains two digital filters, an adder, a threshold block, a unit for setting weight coefficients, a single signal source, a division unit, which serves to isolate a constant component of current or voltage in a transient process.

The disadvantage of this device is the inability to allocate alternating current or voltage against non-fluctuation interference.

From the prior art [Treichler J.R. et al. A new approach to multipath correction of constant modulus signals / IEEE Transactions on Acoustics, Speech, and Signal Processing, 1983, V. 31, N. 2, pp. 459-472] there is a known adaptive filter that suppresses non-fluctuation interference, focused on the reception of signals with a constant envelope. It contains the first and second delay lines, consisting of series-connected delay elements, multipliers of the output signals of the first delay line with weight coefficients, the outputs of which are connected to the inputs of the adder, the output of which is the output of the device and at the same time, the input of the adaptive adjustment of weight coefficients, consisting of a block computing a module, a subtracter, a gain block, multipliers with the output signals of the second register and integrators.

In the specified device to configure the vector of weights of the non-recursive filter, a method is used that implements the procedure:

where d is a coefficient determining the degree of inertia and stability of the adaptation process; (⋅) * is the sign of complex conjugation.

In this case, the objective functional M {ƒ _i } is minimized, where ƒ _i = (| y _i | ² -1) ² ; M {⋅} is the sign of statistical averaging. As a result of AF operation, the envelope of the received process is aligned and, as a result, the influence of interference is reduced.

The disadvantage of this device is the inability to suppress non-fluctuation interference, which, like the useful signal, has a constant envelope.

The closest technical analogue (the prototype of the proposed utility model) is an adaptive filter containing the first and second delay registers, consisting of series-connected delay elements, multipliers of the output signals of the first delay register with weight coefficients, the outputs of which are connected to the inputs of the adder, the output of which is the output of the device and at the same time - the input of the adaptive adjustment circuit of the weight coefficients, consisting of a subtractor, amplification unit, multipliers the output signals of the second delay register and integrators [B. Widrow, S. Stearns Adaptive Signal Processing / Trans. from English - M.: Radio and Communications, 1989 .-- 440 p., Fig. 1.24b].

In the closest analogue, a method that implements the procedure is used to fine-tune the vector of weight coefficients of a non-recursive filter:

W _{i + 1} = W _i + 2dε _i X _i ,

where d is a coefficient determining the degree of inertia and stability of the adaptation process; ε _i is the error signal.

In this case, the mean square error ε _i is minimized. As a result of the adaptive filter (AF), the effect of interference on the signal is reduced.

The disadvantage of the prototype device is the low accuracy of noise suppression and the inability to suppress several non-fluctuation, for example, harmonic interference.

Disclosure of the essence of the utility model.

The technical problem to be solved by the claimed utility model is to increase the accuracy of filtering non-fluctuation interference, taking into account known information about the phase structure of the signal (phase pulse shape), different from the phase structure of the interference.

The technical result that is achieved in the proposed utility model is to reduce the level of non-fluctuation interference at the output of the adaptive digital filter by providing the possibility of implementing the method of tuning the vector of weight coefficients taking into account the known information about the phase structure of the signal (phase pulse shape), as a result of which the suppression non-fluctuation interference having a phase structure different from the signal, and the filtering accuracy is increased.

The stated technical problem is solved and the technical result is achieved by the fact that an adaptive digital filter for suppressing non-fluctuation interference, containing the first and second delay registers, consisting of series-connected delay elements, multipliers of the output signals of the first delay register with weight coefficients, an adder, multipliers with the output signals of the second register delay, integrators, according to the claimed utility model, includes a converter of the complex envelope of the signal in the complex o conjugate envelope, the input of which is the input of the device, and this input is connected to the input of the first delay register, consisting of series-connected first, second, third delay elements, the outputs of which are connected to the first inputs of the first, second, third multipliers of the output signals of the first delay register with weighting factors, while the outputs of these multipliers are connected to the inputs of the first adder, the output of which is the output of the device and at the same time - the input of the adaptive circuit adjustment of weighting coefficients, containing a series-connected derivative calculation unit, a module calculation unit, a second adder, to the second input of which a constant voltage is applied - G proportional to the derivative of the phase pulse of the signal, and the output is connected to the input of the fourth multiplier, the second input of which is connected to the output of the first adder and the third input is designed to supply a constant voltage - d, which determines the degree of inertia and stability of the adaptation process, and the output of the fourth The amplifier is connected to the first inputs of the fifth, sixth, and seventh multipliers, the second inputs of which receive the output signals of the corresponding series-connected fourth, fifth, and sixth delay elements of the second delay register, while the input of the second delay register is connected to the output of the complex envelope signal into a complex conjugate envelope and the outputs of the fifth, sixth, seventh multipliers with the output signals of the second delay register are connected through the first, second and third integrators with about the second inputs of the first, second and third multipliers of the output signals of the first delay register with weights.

A brief description of the drawings.

The essence of the utility model is illustrated by the following drawings:

in FIG. 1 presents a structural diagram of the inventive adaptive digital filter to suppress non-fluctuation (non-noise, structural) interference;

in FIG. 2 shows the pulse and amplitude-frequency characteristics of the inventive adaptive digital filter under the action of harmonic interference;

in FIG. Figure 3 shows the dependences of the probability of error when receiving a signal against a background of harmonic interference;

in FIG. 4 presents the dependences of the probability of error on the amount of harmonic interference when receiving a signal using the inventive adaptive digital filter;

in FIG. 5 shows a plot of the probability of error on the length N of the inventive adaptive digital filter.

The drawing in FIG. 1 contains the following items:

1 - Converter complex envelope of the signal in a complex conjugate envelope;

2, 3, 4 - the first, second, third delay elements of the first delay register;

5, 6, 7 - the first, second, third multipliers of the output signals of the first delay register with weights;

8 - the first adder;

9 - block computing the derivative;

10 - unit calculation module;

11 - second adder;

13, 14, 15 - the first, second and third integrators;

12 - the fourth multiplier,

16, 17, 18 - the fifth, sixth, seventh multipliers with the output signals of the second delay register;

19, 20, 21 - the fourth, fifth, sixth delay elements of the second delay register.

Implementation of a utility model.

An adaptive digital filter of non-fluctuation interference (Fig. 1) contains a converter of the complex envelope of the signal into a complex conjugate envelope 1, the input of which is the input of the device, and this input is connected to the input of the first delay register, consisting of the first, second, third delay elements 2 connected in series, 3, 4, the outputs of which are connected to the first inputs of the first, second, third multipliers of the output signals of the first delay register with weights 5, 6, 7. The outputs of these multipliers 5, 6, 7 are connected to the inputs of the first adder 8, the output of which is the output of the device and at the same time is the input of the adaptive adjustment of weight coefficients circuit containing the derivative 9 calculation unit, module 10 calculation unit, the second adder 11, the second input of which is supplied with a constant voltage - G, proportional to the derivative of the phase pulse of the signal, and the output is connected to the input of the fourth multiplier 12, the second input of which is connected to the output of the first adder 8, and is constantly fed to the third input e voltage - d, which determines the degree of inertia and stability of the adaptation process. Moreover, the output of the fourth multiplier 12 is connected to the first inputs of the fifth, sixth, seventh multipliers 16, 17, 18, the second inputs of which receive the output signals of the corresponding series-connected fourth, fifth, sixth delay elements 19, 20, 21 of the second delay register, while the input the second delay register is connected to the output of the complex envelope envelope converter into the complex conjugate envelope 1, and the outputs of the fifth, sixth, seventh multipliers with the output signals of the second delay register 16, 17, 18 connected through the first, second and third integrators 13, 14, 15 with the second inputs of the multipliers of the output signals of the first delay register with weights 5, 6, 7.

The operation of the device is based on the use of known data on the shape of the phase pulse.

We use the signal recording through the complex envelope:

s (t) = Re [A (t) ⋅exp (jω ₀ t)],

where A (t) = A ₀ exp [-jΨ (t)] is the complex envelope; Ψ (t) is the information component of the signal phase.

Define the modulus of the derivative of the complex envelope:

| A '(t) | = A ₀ | -jΨ' (t) ⋅ exp [-jΨ (t)] | = A ₀ | Ψ '(t) | = A ₀ ⋅ G,

where G = | Ψ '(t) | - the module of the derivative of the information component of the phase (slope of the phase pulse) of the signal.

The inventive device implements a method of adjusting the vector of weights, controlling the shape of the phase pulse of the received signal, and taking into account equation (I) has the following form:

With its help, the target functionality is minimized.

where y _i are discrete samples of the complex envelope A (t); M is the sign of mathematical averaging.

Weighting factors W _i are formed by the first, second and third integrators 13, 14, 15.

The input discrete samples of the complex envelope x _i go to the input of the first delay register, consisting of first, second and third delay elements 2, 3, 4, connected in series, and from their outputs go to the inputs of the multipliers of the output signals of the first delay register with weighting factors 5, 6 , 7 sample vectors X _i with weighting coefficients W _i generated by the first, second and third integrators 13, 14, 15. Signals from the outputs of the multipliers of the output signals of the first delay register with weighting factors 5, 6, 7 are applied to the inputs of the first adder 8, the output of which is the output of the device. This structure is a non-recursive filter that implements a weight vector adjustment procedure that controls the shape of the phase pulse of the received signal and generates samples y _{i of the} output signal of the adaptive digital filter.

In the process of adaptation, the correction of the weight coefficients W _i occurs in accordance with equation (II). For this, the output signal of the adaptive digital filter is differentiated by the derivative calculation unit 9, then the derivative module in the calculation unit of the module 10 is calculated, and the signal is supplied to the second adder 11, on. the second input of which is supplied with a constant voltage - G, proportional to the derivative of the phase pulse of the signal. This allows you to compare the current value of the slope of the phase pulse of the signal with the ideal a priori specified value G.

If there is non-fluctuation (non-noise, structural) interference in the communication channel, these values will differ, and the second adder 11 will generate a control voltage proportional to their difference, which will be used to adjust the vector of weight coefficients in accordance with equation (II). To this end, the signal from the output of the second adder 11 is fed to the input of the multiplier 12, the output signal of the adaptive filter y _i is supplied to the second input, and a constant voltage d is applied to the third input, which determines the degree of inertia and stability of the adaptation process. The signal thus formed is multiplied in the multipliers with the output signals of the second delay register 16, 17, 18 with the sample vector of the complex conjugate envelope X _i * delayed in the second delay register on the delay elements 19, 20, 21:

The output signals of the multipliers with the output signals of the second delay register 16, 17, 18 are fed to the inputs of the first, second and third integrators 13, 14, 15, which carry out the correction of the vector of weight coefficients W _{i + 1} .

In FIG. 2 shows the impulse response of the claimed device in steady state under the action of harmonic interference for the length of the first and second delay registers N = 256 elements, as well as the corresponding amplitude-frequency characteristic. The real Re and imaginary Im parts of the coefficients w _k are shown separately. AF in the process of adaptation generates a “zero” amplitude-frequency characteristic at the frequency of harmonic interference, that is, performs rejection of the interference.

In FIG. Figure 3 shows, for example, the dependences of the probability of erroneous reception of a signal against a background of harmonic interference with a relative intensity μ. Dashed lines correspond to the claimed device, solid lines are constructed without an adaptive filter. The gain from using the claimed device in terms of signal to noise ratio reaches 2 times (3 dB).

In FIG. Figure 4 shows the dependences of the probability of error on the number of harmonic noise simultaneously present at the input of the receiver and distributed with a uniform step in the band of 1.5 π / T. The device allows you to effectively deal with at least 4-5 interference.

In FIG. 5 are graphs illustrating the dependence of the probability of error on the length N of an adaptive digital filter (the number of delay elements of the first and second delay registers). With increasing N, the probability of error decreases, stabilization occurs at N≥256.

Claims (1)

An adaptive digital filter for suppressing non-fluctuation interference, containing the first and second delay registers, consisting of series-connected delay elements, multipliers of the output signals of the first delay register with weight coefficients, an adder, multipliers with the output signals of the second delay register, integrators, characterized in that it includes a converter of the complex envelope of the signal into a complex conjugate envelope, the input of which is the input of the device, and this input is connected to the input m of the first delay register, consisting of series-connected first, second, third delay elements, the outputs of which are connected to the first inputs of the first, second, third multipliers of the output signals of the first delay register with weighting factors, while the outputs of these multipliers are connected to the inputs of the first adder, output which is the output of the device and at the same time - the input of the adaptive adjustment scheme of the weight coefficients, containing a series-connected production calculation unit No, the module calculation unit, the second adder, to the second input of which a constant voltage is applied - G proportional to the derivative of the phase pulse of the signal, and the output is connected to the input of the fourth multiplier, the second input of which is connected to the output of the first adder, and the third input is used to supply constant voltage - d, which determines the degree of inertia and stability of the adaptation process, and the output of the fourth multiplier is connected to the first inputs of the fifth, sixth, seventh multipliers, to the second inputs of which the output signals of the corresponding fourth, fifth, and sixth delay elements of the second delay register are received, the input of the second delay register being connected to the output of the complex envelope converter to the complex conjugate envelope, and the outputs of the fifth, sixth, and seventh multipliers connected to the output signals of the second delay register through the first, second and third integrators with second inputs of the first, second and third multipliers of the output signals of the first register arms with weights.

Images